Growth Product Design: Engineering the Ad-to-Landing-Page Funnel

$1.61 revenue per session

Winning funnel vs ~$0.02 on fear angles

2.23% CVR

Winning funnel vs 0.02-0.16% on fear angles

94% of revenue

From 35% of traffic

Live in-market test — figures reflect data through Jul 14, 2026 and update as the experiment matures.

Moved paid-social traffic off product pages and onto purpose-built landing pages — pairing each ad angle to a matched page and offer, then letting conversion data pick the winner.

Focus Areas:

Funnel architecture
Landing page design
Offer & buy-box design
Ad-to-page message-match
Experiment design & analysis

Problem

Paid social drove clicks straight to product detail pages built for organic and returning shoppers. Cold, skeptical traffic landed on a dense catalog page with no single narrative — and bounced. The mandate: capture that traffic on dedicated landing pages engineered around one promise, one offer, and one clean path to cart.

This was a team effort. Joe Nappi built the paid static and video ads. I owned the funnel design — landing-page architecture, offer and buy-box design, and the ad-to-page message-match — and ran the performance analysis that steered spend. Rose partnered on production builds, Danny set buy-box strategy, and Sarah coordinated UGC creators.

Cross-functional Setup

The Uphill Battle

This wasn't a clean launch. Growth+ went to market while ProClinical — an adjacent product in the same line — was already running paid ads, so the two competed for the same audience and budget, and reported CPAs were inflated by cross-attribution between them. Watching real sessions in Microsoft Clarity, I caught the mechanism: a shopper would land on a Growth+ page, choose a bundle, and add to cart — only for the cart to already contain ProClinical from an earlier visit. The products were persisting across sessions and colliding at checkout, cannibalizing each other instead of complementing. That reframed the goal — the funnel didn't just need a better page, it needed to separate Growth+ from ProClinical in the shopper's mind. And it reframed the metrics: with CPA cross-contaminated at the account level, I steered on page-level measures — conversion rate and revenue per session — that the clash couldn't distort.

Funnel Evolution

1. Foundation — PDP Visual System

Before the funnel, I built the product's visual language: every below-the-fold tile and lifestyle image on the live product page, plus the site hero images. (Copy and Shopify layout by the Shopify manager.)

MyBiotin Growth+ product page — full desktop view of the below-the-fold tiles and lifestyle imagery I designed

MyBiotin Growth+ product page — continued below-the-fold sections

2. Concept 1 — The Sliding-Offer Landing Page

First concept: a single-scroll, cost-led page ("Thicker hair — for less than $1.50/day") where the offer card tracked alongside the content as you scrolled. Built end-to-end as one HTML block — authored with Claude and embedded in Replo, then hand-tuned for mobile responsiveness. Killed before launch — the sticky-sidebar pattern couldn't survive a modular, reorderable section system.

Concept 1 landing page — sliding-offer layout with the "$1.50/day" cost-led hero

Concept 1 landing page — full scroll showing the offer card tracking alongside the content

3. Concept 2 — Top-Anchored Offer, Sliding Hero

Second concept: the offer confined to a single buy box at the top ("…for less than $1.17/day"), editorial sections below. Also shelved pre-launch — but it settled the pattern that carried forward: offer at the top, one buy box, full-width story sections beneath. Same build path — HTML authored with Claude, embedded and made responsive in Replo.

Concept 2 landing page — top-anchored buy box with the "$1.17/day" offer

Concept 2 landing page — full scroll with editorial sections beneath a single buy box

4. The Proven Format — The Listicle Test

Pivoted to a battle-tested listicle format and turned it into a controlled experiment. Six pages: three for a younger persona, three for an older persona, each mapped to a distinct angle — price / value, anti-shed, follicle-reset, hormonal. The winning pages carried a price-targeted headline and a matching price-targeted closer. Copy was formulaic by design (written fast, kept compliant); the story was carried in the copy, with imagery built in Figma or generated in Higgsfield. This was the shift to building natively in Replo — I assembled and shipped the pages in-platform, with Claude providing only the copy (light art direction from me) and the buy-box component.

Listicle test landing page — long-form "10 reasons" article format, full desktop view

Listicle test landing page — second variant, full desktop view

1A Younger ·
Ponytail / Moms

2A Older ·
Shedding Panic

1B Younger ·

Follicle Reset

2B Older ·
Hormonal / No-Rx

1C Younger ·
Price vs. $88 Brand

2C Older ·
Anti-Shed + Value

Listicle 2C — winning page, offer and buy box

5. The 1:1 UGC-Matched page

The newest page closes the loop the data opened. In the listicle test, Stacey's UGC video pulled the strongest hook of any creator (2.79% CTR) — but it was stranded on the worst-converting page (2A, 0.04%). The fix wasn't a new hook; it was a page built to match it. So the 21-Day Challenge page ("Thicker hair — for less than half the price") was designed as two matched versions — one led by Stacey, one by Kyra — each opening with its creator's own footage, so the ad and the landing page are the same person, the same promise, the same imagery, top to bottom.

This is the evolution the whole case study builds toward: the listicle proved message-match on the angle; this page takes it all the way to a true 1:1 ad-to-page flow — creative continuity, not just copy continuity. Built and pending review; results not yet claimed. On tooling, it closed the arc: mocked with Claude, then built end-to-end natively in Replo — everything but the reusable buy-box component.

Stacey's UGC ad — creator video in a phone frame; the test's strongest hook at 2.79% CTR

Stacey's paid-social ad as it ran on Meta

21-Day Challenge landing page — the 1:1 UGC-matched page (desktop)

21-Day Challenge landing page — full scroll

Kyra's UGC ad — creator video in a phone frame

Kyra-led 21-Day Challenge page — hero opens with Kyra's own footage to match her ad

The Data

Two-layer split: The ad's only job is a cheap click (CTR). The page + offer's job is to convert it (CPA). Separating the two exposed the real lever — and it wasn't hook strength.

The strongest hooks didn't convert.
Click-through rate by ad, sorted high to low. Color shows which landing page each ad drove.

The value angle earned far more per visitor.
Revenue per session by landing-page angle.

Confirmed at the register: The decisive cut wasn't on-page CVR — it was Shopify orders. Only the two value/price pages produced real orders; the four fear- and shed-angle pages produced $0 between them. (Replo's broad attribution model had inflated revenue roughly 3× over order-level truth.) The value angle wasn't just the best performer — it was the only one that closed.

Two winning pages, two different jobs: Order-level data showed the anti-shed page (2C) skewed new (76%) — a cold-acquisition engine — while the price page (1C) skewed returning (61%) — a warm / upgrade page. Same category, two distinct roles.

Takeaway

The lever wasn't a louder hook or a scarier fear — it was message-match. Carrying one promise (thicker hair, for less than half the price of the $88 brand) intact from ad → headline → closer → offer is what converted. Growth Design isn't decoration on a page; it's engineering the whole path so the promise never breaks.